Mr. Schönegger, the history of the PLC goes back to 1968. Interestingly, Richard E. Morley - one of the two fathers of the PLC - was particularly opposed to calling the PLC a 'computer' at the time. Morley felt that such a designation would jeopardize acceptance among control specialists.
How would you describe the historical development of the PLC from today's perspective?

Stefan Schönegger: Until the late 1990s, a PLC was actually closer to a rigid hardware function than a flexible computer. However, this changed abruptly with the technological developments surrounding the so-called soft PLC. This expansion was not only limited to the further development of the software, but also in the way consumer hardware - keyword x86 - was used in conjunction with real-time operating systems.

At B&R, we were already offering our customers PLC tasks, visualization and drive control in parallel on one 'computer' in the early 2000s. In doing so, we took on a pioneering role that was later imitated many times over. Both in the form factor of a PLC with the classic onboard IOs and in the form of an Automation PC, i.e. the combination of an industrial PC with an automation system.

The latest trend is to increasingly depict the PLC as software. Do you think the hardware PLC is about to be replaced?

I think it is necessary to differentiate between the two terms hardware PLC - hardware and PLC.
Hardware in the sense of a computer with CPU and memory will always be necessary. Likewise, a PLC in the sense of a function that can process any programmed tasks within the necessary framework conditions - keyword real-time.
What will change is that the combination and selection of the two topics will become increasingly flexible in the future. Whether a PLC is available as a function on a server, IPC or an embedded system will become more than ever a degree of freedom that the customer can decide for themselves.

Stefan Schönegger, responsible for the Controls division at B&R

© B&R

At this year's Hannover Messe, there was a lot of talk about mapping a PLC as a purely virtual solution in the future. What do you think of this trend?

This is the logical next step and is actually just an evolutionary development of today's solution with a soft PLC. This includes the possibility of operating such a PLC directly in a public cloud. However, you have to weigh up very carefully for which areas of application such an approach is or is not suitable. A PLC that is primarily used for IO monitoring with cycle times in the seconds range without real-time requirements is certainly an ideal use case for a virtualized PLC in the cloud. The other extreme is a highly dynamic machine with many servo drives that need to be controlled with µs precision. This requires very tightly meshed real-time behavior on the control system - which is currently not yet conceivable with a virtual controller. But of course there will also be hybrid systems that map some tasks in virtualized form. One example of an application would be a local data center that executes real-time processes very close to the actuators/sensors.

"There will be no either-or" - continued

Compact controller from B&R

© B&R

What new possibilities does virtualization open up?

The major advantages of virtualization technology are a very simplified update scenario and hardware independence. Switching from a PLC-based solution to an industrial PC and back is already possible at B&R today. Complete virtualization also makes it possible to switch to the cloud. The customer no longer has to commit to one of the two worlds at the start of development.

"The major advantages of virtualization technology are a very simplified update scenario and hardware independence"

What division of the 'control' task do you see? Everything in the cloud, some in the edge, some directly in the sensor/actuator?

There won't be an either-or, but - depending on the task - flexible load balancing. This increases the flexibility and also the performance of the systems, but will also lead to more complexity. In future, one of the core tasks of automation specialists will be to provide customers with these benefits without them having to deal with them explicitly. The distribution of intelligence should therefore be fully automated using clever algorithms - of course always with the option of an expert mode to influence the system behavior in a targeted manner.

To what extent do you see the topic of control merging more closely with other disciplines?

This trend will continue. In addition to the familiar topics of motion control, HMI, IO and safety, there will be new topics such as data analysis using AI, for example for vision systems or predictive maintenance. In the future, control and regulation can be imagined as an app that can be combined with any number of tasks - just as we already know it from the consumer world. However, there will be limits the higher the real-time requirements become.

What new communication technology requirements will arise with virtualization?

One important point is the necessary merging of real-time communication and service tasks on a convergent infrastructure. Ultimately, everything revolves around data and therefore the requirement to use it to describe processes semantically correctly and completely. We need to move away from bits and bytes towards usable information. We see all of this with OPC UA and TSN, which is currently being finalized in the standardization committees as OPC UA FX. I have no doubt that the future belongs to these technologies.

To what extent will or must traditional control system manufacturers change? How do you expect the industry to develop?

The much-described convergence of OT and IT will no longer be an issue in a few years' time, but will be a reality in automation technology. The focus will increasingly be on software and digital services. Those who do not manage this change will fall by the wayside.